AI without insight is just noise, so what's the magic formula?

Advertisers and marketers like to say they’re data-obsessed. But is that actually the reality? And how will the rise of AI change that? To answer that question, we interviewed 550 business professionals in the US throughout the summer of 2025.

AI for AI’s sake is nobody’s friend. Used ineffectively, it’s a quick path to banal - or the “slop”, as one leader recently put it to me. If outputs aren’t grounded in reality, speed becomes noise. That’s why the data that feeds your AI tool matters more than ever.

Generic AI “slop” is a huge business risk

AI has a powerful role to play. Right now, our research shows that as many as 50% of business decisions - spanning all use cases and workflow stages - are being made without consumer insights. That’s despite an overwhelming majority of professionals saying insights are valuable and lead to better outcomes. The data we’ve gathered shows that deploying consumer insights drives materially higher ROI, too. 

So, why the gap? Ultimately, it’s all about friction; accessing insights has historically been slow, expensive and/or complex. If insights aren’t embedded directly into the process, many people default to intuition or internal data. 

To solve for this, many companies have been turning to AI tools. Within the advertising and marketing industry, 85% are already using them, with ChatGPT, Gemini, and Copilot leading the pack. 

But here’s the challenge: most AI tools are powered by generic models, without access to verified consumer data. That means a reliance on generalization and web scrapes, as well as a looming risk of hallucinations. They’re providing seamless, quick results - but are they offering genuine, actionable insights? Or in fact, the best summaries of the mediocre (and at times inconsistent) data available on the open web?   

Historical snapshots on the open web just won’t cut it

The best insights live in the strange, unpredictable contradictions which are so inherent to humans. Yet AI tools – without the right data – will strive to give answers which are expected, obvious, or so diluted that they’re ineffective. Think, for example, of an AI-generated media plan which recommends investing in every single channel, or the creative which reflects the most stereotypical assumptions about an audience. Cristina Lawrence, EVP of Consumer & Content Experience at Razorfish, believes this is a systemic challenge.

“The model out of the box is not going to get us there. It’s going to tell you what you’re asking for and it’s just going to make assumptions based on the context of your statement,” she explained.

“It actually takes a fair amount of work using human insights, survey data, and behavioral data to really get to what it is a person is actually thinking.”

Humans are complicated and messy. They’re also contradictory, at several times surprising, and are prone to changing their views and behaviors. If we’re going to understand them, then relying on the static insights, pre-defined audiences and historical snapshots available across the open web just won’t cut it. 

You need to know that the data you’re looking at reflects the audience you actually want, rather than the closest proxy that can be found. You need to be confident the data shows you how they’re feeling and behaving now - not several quarters or years ago. 

The magic formula: Agent, MCP, and the right data

At GWI, we’re all-in with our data being connected to AI tools and agents through the MCP - which has quickly become the foundational standard for how AI connects with data, tools, and workflows. Think of it as akin to how a USB-C provides a universal port for devices; the MCP provides a uniform interface for AI applications to plug into data and services. It gives any user the ability to draw dynamic insights from structured, verified, up-to-date data like GWI’s. And the results are transformational. 

Don’t just take my word for it, though; AI tools make the same assessment by themselves. Whether it’s campaign planning, audience profiling, creative development, or any other use case, we’ve given the industry’s leading tools the exact same prompts, but asked them to produce one output with GWI data vs one without. 

The verdict from one of these tests is typical. The AI tools themselves tell us:

“The GWI output would likely achieve 2-3x better performance because it targets the right people, on the right platforms, with the right messages, rather than relying on general assumptions, on statistics that may not apply to the specific audience, and on generic usage patterns”.

True intelligence isn't artificial - it features human judgement

We believe the most resilient companies are the ones that combine machine intelligence with human understanding. The combination of agent, MCP + GWI is what allows robust, scaled, dynamic consumer insights drawn from over a million people and 50+ markets to be integrated within workflows.

It makes consumer insights more usable and available to any end-user. It makes consumer insights actionable within the moments and workflows where they’re required. 

It also speaks to exactly what professionals are looking for. Our research shows that trust in AI agents is most dependent on consistent accuracy over time, transparent & cited sources, as well as clear explanations of how insights were reached. All of that’s built-in, here.

But advertisers and marketers also want to bring their own experience and intuition to the table. So here’s the beautiful thing: our research verifies that professionals utilizing AI tools are just as likely to turn to their own judgement as non-users; AI isn’t removing human judgment and creativity, but allowing people to make decisions based on much richer, data-fuelled insights. It’s broadening the evidence base and speeding up interpretation, with experience providing the grounding that keeps decisions practical and relevant. 

That combination - structured external data, accelerated by AI, anchored in human judgement - is where the real impact lies and where trust is able to cut through. And while speed and productivity gains are the most obvious early benefits felt by adopters of AI agents, it’s telling that seasoned, heavy adopters are more likely to cite the most valuable gains as better alignment across teams, deeper customer understanding, and unexpected insights. 

This is the sophistication gradient: efficiency is the hook, but transformation is the prize. The table-stakes benefits may be there, but the real impact runs much deeper.

For personalization at scale, you must plug consumer insights into your workflow

GWI has long been known for representative human data at scale across more than 50 markets. That means we’re especially well placed to integrate with AI tools and become a currency for agentic workflows.

You can speak to your target audience in real time, using insights that reflect what people actually think, feel, and do. Through our APIs, you can plug GWI directly into your workflow — defining any audience from the tens of thousands available, and receiving a detailed, text-based summary you can use to shape messaging, train models, or build avatars on the fly.

We’re also seeing growing traction in the probabilistic space. Many clients have traditionally preferred deterministic matching, but that approach can be slow, complex, and weighed down by privacy or legal barriers. By combining the demographic data you already know with GWI’s attitudinal insight, you can uncover what really resonates — then build content or experiences that land in the moment. It’s personalization at scale, in a way that actually works. 

Talk to us

The world's biggest brands and best agencies use our data to power their GTM strategies. But don't just take our word for it. Our G2 page is full of customer reviews, and you can check out our case studies if you want to see more use cases. If you're ready to see the agent + MCP combination in action, you can book a demo today.